SFEBES2017 Poster Presentations Bone and Calcium (27 abstracts)
Wolverhampton Diabetes and Endocrine Centre, New Cross Hospital, Wolverhampton, UK.
Background: Management of PHPT includes successful localization of parathyroid adenoma(s) by nuclear imaging and ultrasound scans allowing focused or minimally invasive surgery. Our aim was to assess the accuracy of parathyroid imaging in surgically-treated patients with primary hyperparathyroidism (PHPT) by correlating the imaging findings with surgical findings and cure of hyperparathyroidism. We also compared the surgical outcomes in patients with concordant and discordant imaging.
Patients and methods: We performed a retrospective review of 92 patients who were operated for PHPT between 2012 and 2016. Mean age 62±24 years, calcium 3.37±0.9 mmol/l, PTH 27.47±24.39 pmol/l. 82 patients had both ultrasound and 99mTc-MIBI/123I subtraction SPECT/CT scans. Findings on both scan - individually and together - were compared with surgical findings and outcomes using pre-defined criteria.
Results: 50(62%) patients had concordant scans and 32(38%) had discordant results. Of the latter, in 27(31%) patients one or both scans failed to localise an adenoma and in 5(6%) patients adenoma was localised on opposite sides. Cure rate was 84% and 94% (P=ns) in patients with discordant and concordant imaging respectively. Patients with discordant scans had lower calcium, lower PTH and smaller adenoma, which weighed less as compared to those with concordant scans (P<0.05 for weight, P=ns for rest). The sensitivity and positive predicted value (PPV) of radionuclide (81% and 90%) and ultrasound scans (76% and 88%) were comparable, while concordance imaging had a sensitivity of 66% with PPV of 96%.
Conclusion: Sensitivity and PPV of US and RN scans was comparable and in line with published values in the literature. Discordance rate was high and in a majority of patients it was due to non-localisation. Cure rate for patients with discordant scans was lower than in patients with concordant scans, but the difference was not statistically significant, possible due to small numbers.